The on-board intelligent system is utilized to identify and warn drivers of fatigue during driving,which can effectively improve traffic safety.Aiming at the problem of low accuracy for small target detection under poor cab lighting,driver fatigue detection algorithm based on YOLOv5-CC is proposed.This algorithm improves the C3 mod-ule structure of Backbone in YOLOv5 by adding convolutional layers and coordinate attention to the module,and combines Backbone and Neck through criss-cross attention mechanism to enhance the model's global feature extrac-tion ability.The experimental results show that the precision of YOLOv5-CC algorithm is 98.0％,the recall rate is 98.3％,the average precision is 98.9％.Compared with other algorithms,the mean average precision is 98.9％,the experimental results are superior.
万方数据
维普
